[lkml]   [1996]   [Jun]   [26]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: Drawbacks of implementing undelete entirely in user space
Alan Cox <> wrote:
>> > That implies that deleting files could potentially increase the space
>> > requirement on the disk (not good). Creating a directory is an "expensive"
>> > operation (uses non-negligible space and cpu/wallclock time).

>> > I don't think those are desirable properties of an unlink() operator.

>> Indeed, because a single line of extra code in ext2's debugfs makes
>> undelete possible as long as the file hasn't been physically overwritten
>> yet.

>Actually I beg to differ given the small size of the patch to keep old
>files around and how trivial it is to switch it on or off, I'll happily
>trade a few microseconds per unlink in user directories for not having to
>fish files off tape.

I think we are talking about different things. What I meant to say
in the first quote above (the second wasn't mine), is that the
undelete feature as such is excellent. However, what is not so
excellent is that if in the .wastebasket directory there will be
created *extra* directories for every file deleted.
Stephen R. van den Berg (AKA BuGless).

"ICMP: The protocol that goes PING!"

 \ /
  Last update: 2005-03-22 13:37    [W:0.040 / U:1.252 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site